Dear FreeSurfer team,

I'm using fsfast to analyze 1.2mm^3 7T BOLD data collected with a partial coverage over the hippocampus and superior temporal lobe.

I ran preprocessing with the following flags:

preproc-sess
-d /functional-data
-s SUB01
-fsd bold
-surface SUB01 lhrh
-fwhm 2
-init-header
-bbr-int /functional-data/SUB04/bold mT2_WB
-per-session

And then I ran the GLM analysis (stimulus-vs-baseline) using selxavg3-sess, keeping the data in the subject's native space (using the '-native' flag combined with '-per-session').

The resulting t-map shows a diagonal line of spurious activation along the edge of the functional volume. Is there a way to eliminate this artifact? (see figures attached)

Moreover, is there a way to use the -native flag together with -per-run?

I want to simultaneously measure the hippocampal activation (in native space) and cortical activity smoothed on the surface; but I want all the data to be processed the same way (-per-run), and currently the volume analysis forces me to use -per-session.

I haven't encountered any previously documented issues similar to this one.

Please let me know if you have any thoughts.

1) FreeSurfer build stamp: freesurfer-darwin-macOS-7.3.2-20220804-6354275
2) Platform: Mac Studio M1 (macOS)
3) selxavg3-sess log: attached
4) analysis config file: attached

Thanks,

Itzik